Automated DB2 Restart Question

Luke (ECS) McGee

Automated DB2 Restart Question
I am interested in the various procedures that might be in use for the automated restart of DB2 after an abnormal termination. I am assuming that some sort of Automatic Operator Software is used to monitor system messages and then take action upon the failure of a DB2 System. There are obviously many details involved with this issue depending on the circumstances of the failure, so maybe we could discuss the issue at a more generic level such as a "must complete" type of failure which stops DB2, by design...

Thanks.



Lee Hayden

Re: Automated DB2 Restart Question
(in response to Luke (ECS) McGee)
Luke,

You should look into a feature of OS/390 called
"AUTOMATIC RESTART MANAGER". There is some information
in The DB2 Data Sharing documentation, and more in:

OS/390
MVS Programming:
Sysplex Services Guide
Document Number GC28-1771


-----Original Message-----
From: McGee, Luke (ECS) [mailto:[login to unmask email]
Sent: Friday, January 05, 2001 10:42 AM
Subject: Automated DB2 Restart Question


I am interested in the various procedures that might be in use for the
automated restart of DB2 after an abnormal termination. I am assuming that
some sort of Automatic Operator Software is used to monitor system messages
and then take action upon the failure of a DB2 System. There are obviously
many details involved with this issue depending on the circumstances of the
failure, so maybe we could discuss the issue at a more generic level such as
a "must complete" type of failure which stops DB2, by design...

Thanks.



Sue Janowitz

Re: Automated DB2 Restart Question
(in response to Lee Hayden)
We use the automation product OPSMVS. We have a message rule that kicks off
if we get a DSNV086E message (DB2 ABNORMAL TERMINATON). It triggers a REXX
that does the following:

Checks to see if we've had more than 2 abends (for the same subsystem) in 10
minutes (an arbitrary number we decided on) - if so, don't try to start up
again, but send message to console to contact DB2 support.

If more than 2 abends but longer than 10 minutes (we assume it's been up and
working for a period of time), or less than 2 abends:
check that MSTR, IRLM, and DBM1 are all down. We wait up to 6 minutes for
them to come down if they are initially active. If they don't come down,
issue a message to console saying they are still up and to contact DB2
support.

If they are down, issue a REXX that does STARTDB2 and connects/starts
dependent applications, including CICS, IMS, Replication Agent, Changeman.
Also issue a message to console that the subsystem was restarted, and to
notify DB2 support.

Hope this helps.

Sue
Sue Janowitz
DB2 Systems Administrator
New England Financial Information Systems
Boston, MA
[login to unmask email]

> -----Original Message-----
> From: McGee, Luke (ECS) [SMTP:[login to unmask email]
> Sent: Friday, January 05, 2001 10:42 AM
> To: [login to unmask email]
> Subject: Automated DB2 Restart Question
>
> I am interested in the various procedures that might be in use for the
> automated restart of DB2 after an abnormal termination. I am assuming
> that some sort of Automatic Operator Software is used to monitor system
> messages and then take action upon the failure of a DB2 System. There are
> obviously many details involved with this issue depending on the
> circumstances of the failure, so maybe we could discuss the issue at a
> more generic level such as a "must complete" type of failure which stops
> DB2, by design...
>
> Thanks.
>
>
>
>
>